> diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c
> --- a/net/bpf/test_run.c
> +++ b/net/bpf/test_run.c
> @@ -1156,6 +1156,18 @@
> skb->ip_summed = CHECKSUM_COMPLETE;
> }
>
> + if (prog->type == BPF_PROG_TYPE_LWT_XMIT && !skb_dst(skb)) {
> + if (!ipv6_bpf_stub) {
> + pr_warn_once("Please test this program with the IPv6
> module enabled\n");
> + goto out;
Should this set ret to an error code before goto out? At this point
ret is 0 from the earlier convert___skb_to_skb() success path, so the
function returns 0 to the bpf() syscall without ever running the BPF
program.
Every other early exit in bpf_prog_test_run_skb() assigns a negative
errno before goto out, for example:
ret = -EINVAL;
goto out;
ret = -ENOMEM;
goto out;
ret = -ENODEV;
goto out;
Returning 0 here means userspace sees a successful test run that never
actually executed. Something like ret = -EOPNOTSUPP would let the
caller know the test could not proceed.
> + }
> [ ... ]
---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
CI run summary: https://github.com/kernel-patches/bpf/actions/runs/22437468613
AI-authorship-score: low
AI-authorship-explanation: Patch iterated through 7 versions over 3 weeks with
detailed maintainer interaction, consistent with human development.
issues-found: 1
issue-severity-score: low
issue-severity-explanation: Missing error return code on early exit path causes
bpf() syscall to return success without executing the BPF program when IPv6 is
unavailable.
